The preparations of toxin and antitoxin hitherto 

 produced have been far from pure, so that it is 

 impossible to indicate their composition and especi- 

 ally their percentage of active bodies. Under such cir- 

 cumstances we cannot at present hope to obtain any 

 idea of the molecular size of the active bodies by means 

 of direct methods, as determination of the freezing or 

 boiling point or of the vapour-pressure of solutions of 

 toxins and antitoxins. 



We have therefore tried how far we could get by means 

 of an indirect method. By studying the diffusion of the bo- 

 dies in question we are able to obtain some knowleadge 

 of their molecular weight. The velocity of diffusion of 

 gases through a porous wall being inversely proportionate 

 to the square root of their molecular weight, Pickering 

 tried if the same would not be the case for fluids dif- 

 fusing in water. From his values that differ very much 

 from those of other authors he concluded, that the dissol- 

 ved molecules do not behave as gas-molecules. Notwith- 

 standing, this rule that the velocity of hydrodiffusion 

 is in inverse proportion to the square root of the mole- 

 cular weight, has later on been tested by Hiifner and 

 Euler, and has been found right as regards the main 

 points; still it does not apply to ions. It k denotes the 

 velocity of diffusion, m the molecular weight, then at 



3 _^. 



